# Split transform

The Split transform is used to partition input data into two distinct outputs based on a specified condition. This transform evaluates each row of the input data against the defined condition and directs the rows to one of the two outputs accordingly.

![An example Pipeline Builder graph with the Split transform](../../../assets/8a8212cd8b843513fae0bdfb5ca136f1c70a33a4c2c2f836a66b8d2f3c0b2c41.webp)

## Supported pipeline types

:::callout{theme="warning"}
The Split transform is only supported in standard batch (Spark backed) and faster pipelines (DataFusion backed). The **Split** option will not appear in streaming pipelines.
:::

For the current list of supported execution modes, see the [`Split on condition`](../functions-index/index.md#split-on-condition) entry in the function reference.

## Add the transform to your pipeline

Select a single table node in your graph, then use either of the following entry points:

* Select **Split** in the actions that appear beside the selected node.
* Select the **Split table** option in the transform actions toolbar.

![Screenshot of the Split transform in Pipeline Builder](../../../assets/65f4bd35fcb05a56b28752b9ed597e785a752a16cd04a82b1bae638e76e4dfde.webp)

:::callout{theme="neutral"}
If the **Split** option does not appear, confirm that you are working in a standard batch or faster pipeline and that you have selected exactly one table node. The option is hidden when it is unavailable, so it does not appear in streaming pipelines or when the selected node is not tabular, such as a media set or file-based input.
:::

## Condition

The Split transform allows you to define a condition that determines how the input data is divided. This condition is a logical expression that evaluates to either **True** or **False**.

## Outputs

The **True** output will contain rows for which the condition evaluates to true. These rows are directed to the first output.

The **False** output will contain rows for which the condition evaluates to false. These rows are directed to the second output.

To preview the outputs, you can use the dropdown in the top left of the bottom panel. Select **True** to see the True output or **False** to see the False output.

![Screenshot of an example Split transform preview table](../../../assets/caf12c17cb485761b53c64a3eec70ffed4bc1073931feb33b4b950aaae360873.webp)

To use the outputs in downstream transforms, select the transform, and then select either the **True** or **False** output to use as your next input.

![Screenshot of the Select input Split transform window](../../../assets/4ce4d9ab2f4103f9596b8d071843f5d7d1a8666df7e993b2262134ce6ce39ad7.webp)

## Example

Consider a dataset of customer orders. You want to separate orders into two categories: high-value and low-value orders, based on a threshold value.

In this case, the **Condition** will be `order_value > 1000`.

The **True** output will contain orders where `order_value` exceeds `1000`, and the
**False** output will contain orders where `order_value` does not exceed `1000`.

![Screenshot of an example Split transform configuration](../../../assets/40b9a844aa1bab41b561e7879244b5f4ddffc763fe82a430f33fa37773738d3f.webp)